Understand the boundaries before you connect.
Read how Memrain stores knowledge, how context reaches your AI, and which responsibilities stay with the operator.
Know where information goes.
In the documented AWS deployment, stored knowledge lives in your account. Retrieval sends selected passages to the connected AI client, and that client handles them under its provider’s policies. Bedrock processes configured model calls. Default Cloudflare Tunnel ingress passes through Cloudflare; Caddy mode offers a different ingress path.
Use the right access boundary.
Credentials and source grants determine which memory an MCP caller can read or write. Team separation is enforced in the application and remains pilot-grade. The operator admin panel has broader access and is not a personal workspace for every member. Evaluate the documented configuration and limits before adding sensitive information.
Inspect the evidence.
The source, security policy and deployment documentation are public. There is no independent security audit, compliance certification or contractual SLA to download. A security review should start from the actual code and your deployment configuration, not an assurance badge.
The website and your deployment are different.
This marketing website has its own privacy and legal information. It does not host your Memrain knowledge store. Read the existing website policies for website processing; assess your self-hosted deployment and chosen AI providers separately.
Keep the context. Move the work forward.
Connect your knowledge to the AI you already use. Find your starting point and make the next conversation count.